The Indians, who were successively conquered by foreign invaders, spoke hundreds of different dialects. Bancroft estimates that there were six hundred between Alaska and Panama. Ameghino speaks of eight hundred in South America. Most of these, however, are mere derivatives from a single mother tongue like the Aymara and the Guarani. We quote these figures for what they are worth. Philology has no precise definition of what constitutes a language, and any one can add to or deduct from the numbers given according to the point of view from which he considers the matter. As an illustration of this, it may be mentioned that some philologists estimate the languages of North America at no less than thirteen hundred, whilst Squier would reduce those of both continents to four hundred. These dialects present a complete disparity in their vocabulary side by side with great similarity of structure. "In America," says Humboldt, "from the country of the Esquimaux to the banks of the Orinoco, and thence to the frozen shores of the Straits of Magellan, languages differing entirely in their derivation have, if we may use the expression, the same physiognomy. Striking analogies in grammatical construction have been recognized, not only in the more perfect languages, such as those of the Incas, the Aymara, the Guarani, and the Mexicans, but also in languages which are extremely crude. Dialects, the roots of which do not resemble each other more than the roots of the Slavonian and Biscayan, show resemblances in structure similar to those which are found between the Sanscrit, the Persian, the Greek, and the Germanic languages." [42] The fact that the different dialects, or languages, as some call them, "are mere derivatives from a single mother tongue," argues strongly, of course, for ultimate race unity. The following summary of evidences on the substantial unity of race in American peoples is from Marcus Wilson, and will be found valuable: